# [EpiData-list] Issues of recent days

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Fri Sep 29 13:31:35 CDT 2006

```Regarding questions the latest days.
ALWAYS included:
a. Is this epidata entry or epidata analysis
b. Which version of the software are you using (number and country versions)
c. What is the operating system you are using

A few suggestions for the specific problems:
1.

Sara gets :

>
So this is a "fortunate" situation since it is reproducible. I suggest you put your chk and rec file into a zip archive (tools menu) and send to info @ epidata.dk

We will run your file in a controlled environment, since your combinations of questions etc. points to an error or bug in the epidata entry programme.

................

Gerson Zanetta de Lima
have problems of a different nature. The problem is to get the software to "understand" that a simple calculation:

>I can make operations like defining a new floating variable and using the command
>newvariable = height/2 or newvariable = height*height
>
>
But then trying to do the same with "define imc ##" does not work. This
is due to the problem that EpiData Entry and Analysis are more strict in
controlling variable conversions than e.g. Epi6. The problem is that:

>weight/(height^2) or weight/(height*height)
>
>
results in a float resultat - meaning the content cannot end in an
integer. What should the software do with the calculations ending in
e.g. 32.14 or 31.89. So in this situation if you only want the integer
part of the result (here 32 and 31 or rounded 32 in both cases) make an
explicit conversion:

From help file of entry you find the function Integer(x), which
converts x to an integer value.
define imc ##
imc = integer(weight/(height^2))

The same in analysis: example variable "Timer" which is float:
TIMER Float
X Integer 2 0
x = int(timer)
Data type mismatch
x = integer(timer)

Since the function int(timer) returns a float value, but integer(timer)
returns an integer value.

Another example of this more careful value handling is when EpiData
Analysis reads date variables. An illegal date will be rejected, e.g.
"01/01/1", which would be accepted by Epi6.

.......................................................
Regarding Henriettes installation problem:
There is no need to try different language versions. The exe file is
always the same.

I would do exactly this:
1. Install analysis on a different machine
2. Create a folder where you want analysis, e.g. c:\program files\epidata
3. Copy the epidatastat.exe to the "troublesome" machine in that folder.
4. Run the exe file (which will tell you that help files are missing).

After that you can copy the documentation from the subfolders on machine
1 to the folders on the troublesome machine.

regards
Jens Lauritsen
```